Lokad Integration?
I know Patrick brought this up in San Diego, but I have been talking with Lokad about developing support for RetailOps. Is anyone else interested in this functionality.
For those that are not familiar with Lokad, they use machine learning and quantile modeling to forecast inventory purchasing based on hundreds of factors (both on the supply and demand side).
Let me know if anyone is interested in talking about this.
-
Jason:
I have a Lokad account and have done quite a bit of experimentation a couple of years ago. As you have likely realized, their computations are quite complex and the accuracy of the forecasts or safety stock calculations are based on the quality of data being provided. Creating item sales history in Lokad's time series format should be quite easy for RetailOps to put in place. However, to really make the calculations work well, you will need item level lead times and ways to mark items that are currently in a promotion along with other data fields being added to RetailOps and the data in those fields maintained or automated in some way.
I would fully support a Lokad integration, but the pricing for Lokad has evolved a lot over the years. The latest pricing based on annual revenue is not ideal as some merchants will have way different margins than others.We would be looking at probably around $2500 per month and it would take a while to get the calculations full dialed-in. That could pay for itself quickly if we can reduce inventory and I think we could.Once we are happy with the calculations, we would then need the results to be populated back into RetailOps to assist in making purchase order decisions on PO screens. This will work very well for established items, but an experienced buyer will need to make decisions on stocking of new products based on reviewing data from similar items.
Please let me know how I can help. -
Patrick-
As always, I appreciate your input. My primary goal in exploring this integration is not to replace an experienced buyer, but rather to streamline and automate as much of the ordering process as possible. That seems to mesh with the RetailOps mission of providing efficiency through automation to make our jobs easier.
The other option for me is to higher a second buyer, so even on their managed plan, we're coming out ahead, and that's not factoring in any of the cashflow, cost of capital, etc. costs that we would be saving by reducing stock levels.
The other reason for a direct integration is the visibility that will provide. As you mentioned, getting insights into actual lead time (time from when the demand is realized by Lokad until the inventory is actually received in our warehouse) can be tracked for every PO automatically to give us a real glimpse into the delays that can occur at the vendor level or at our level.
I'll keep you posted on our progress as we continue the conversation.
-
Sorry for the lack of updates. We are still in the discovery and setup phase.
In the interest of expediting the deployment we're handling everything outside of RetailOps currently. Hopefully we get to the point down the road where the results justify the effort and RetailOps has the resources to do a proper integration.
I will keep you posted as we make progress and start to get actionable data out of Lokad.
-
It's too early for me to comment on whether or not we're happy, but from our experience to date it seems as though it will provide valuable insight into our inventory planning.
Unfortunately we can't get the full benefit from it without a full integration to RetailOps, which probably won't happen any time in the near future.
I'll provide more feedback at the end of Q4 after we get through that madness, which is when it will really prove valuable.
Please sign in to leave a comment.
Comments
9 comments